1

Does Having Women in Positions of Power Reduce Gender Inequality In Organizations?

A Direct Test

Mabel Abraham mlba@mit.edu

PhD Candidate, MIT Sloan School of Management June, 2013

Draft under review: Administrative Science Quarterly

Please do not cite or circulate without permission of the author.

ABSTRACT

This study revisits the common view that female managers attenuate workplace gender

inequality among employees. While many of the studies that have explored the relationship

between having more women in management and gender inequality have found an association

between the two, these studies have inferred that manager influence is driving this observed

association. Because direct data on the impact of male versus female managers on the career

outcomes of employees is difficult to obtain, prior studies have measured the impact of female

managers on gender inequality using industry or organization-level data. This is problematic

because it is plausible that organizations with more women in management have less gender

inequality not because female managers are having an impact, but rather because the same

mechanisms driving women into management are also reducing gender inequality among non-

managerial employees. Using unique panel data from a large retail financial services firm, I

examine whether female managers contribute to less gender inequality among the employees

reporting directly to them. I find that female managers are more likely to allow their subordinates

to use flexible work arrangements and that female managers reduce gender inequality in terms of

wages, but only for subordinates in the lowest organizational ranks.

Abraham

1

INTRODUCTION

Why does gender inequality in organizations persist? Studies in organization theory and

sociology have demonstrated that organizational decision makers impact gender inequality by

controlling the distribution of rewards to organizational employees (e.g. Reskin, 2000; Elvira and

Graham, 2002; Fernandez and Fernandez-Mateo, 2004; Cohen and Huffman, 2007; Bjerk, 2008;

Castilla, 2008, 2011; Briscoe and Kellogg 2011). In particular, the gender of the organizational

decision maker is posited to play a role, with male managers allocating resources, such as wages,

in a way that benefits other men and disadvantages women (Bridges and Nelson, 1989; Nelson

and Bridges, 1999; Bielby, 2000; Reskin, 2000). The implicit assumption of these studies is that

increasing the prevalence of women in management will lead to the attenuation of gender

inequality in the workplace. And, indeed, other studies have shown that a greater proportion of

women in management is associated with less gender inequality (e.g. Baron, Mittman, and

Newman, 1991; Pfeffer, Davis-Blake, and Julius, 1995; Kulis, 1997; Hultin and Szulkin, 1999;

Cohen and Huffman, 2007; Huffman, Cohen, and Pearlman, 2010).

While existing research has demonstrated this association between women in management and

gender inequality, it has not accurately identified the underlying mechanism. Studies that have

shown lower levels of gender inequality in settings where there are more women in management

have speculated that it is female managers that make the difference. They have inferred that,

because of homophily, in-group preference and a reduction in gender stereotypes, female

managers produce more gender equitable outcomes. For example, past research finding an

association between women in management and gender inequality has suggested that female

managers attenuate gender inequality within organizations by facilitating interpersonal

interactions for female non-managerial employees and reducing the salience of gender (Hultin

Abraham

2

and Szulkin, 1999, 2003). Similarly, other studies attribute this observed association to female

managers evaluating female employees more favorably than do male managers (Shin, 2012).

However, while female managers have commonly been credited with reducing workplace gender

inequality, the evidence in past studies has been indirect. These studies have shown the

relationship between the percent of women in management positions and gender inequality using

industry- or organization-level data. This indirect approach is problematic as there are credible

alternative explanations for this relationship. Specifically, by associating the percent of women

in management with gender differentials among employees, it is impossible to distinguish

whether observed differences in gender inequality are the result of the impact of female manager

decision making or other unobserved organizational processes. For example, it is plausible that

unobserved organizational characteristics may be simultaneously driving the presence of women

in management and gender inequality among employees.

Our limited understanding of the mechanisms driving the relationship between women in

management and gender inequality is, in part, the result of empirical challenges. Directly

identifying the impact that female managers have on gender inequality requires a comparison of

gender inequality outcomes for employees reporting directly to male versus female managers, in

a setting where managers have the power to affect resource allocation. Using unique longitudinal

data from a large retail financial services firm, I address this challenge. By identifying managers

and the employees whom they supervise, I can directly link differences in gender-based wage

inequality, job segregation, and use of flexible work arrangements among employees to the

gender of the manager. I find that female managers are more likely to allow their subordinates to

use flexible work arrangements and that female managers reduce gender inequality in terms of

wages, but only for subordinates in the lowest organizational ranks. By isolating the effect of

Abraham

3

managers on their direct subordinates, this study disentangles the potential effect of female

managers on gender inequality from the effect of other factors related to the organizational

setting. This more direct examination deepens our understanding of whether female managers

attenuate gender inequality by identifying the direct effect of female managers on employee

outcomes.

To analyze the effect of female managers on inequality for their subordinates, I proceed as

follows. First, I discuss existing organizational and sociological theories describing, and

attempting to explain, the relationship between women in positions of power and workplace

gender inequality. Second, I introduce the research setting and analytical method used to test

whether female managers directly reduce gender inequalities in terms of wages, job segregation,

and use of flexible work arrangements. Third, I present results which demonstrate that the impact

of female managers for gender inequality is limited to employees in the lowest organizational

ranks. I conclude with a discussion of the theoretical and practical implications of the research

findings for gender inequality within organizations.

THE PROPOSED EFFECT OF FEMALE MANAGERS ON GENDER INEQUALITY

Given that women now hold over forty-percent of all management positions (Bureau of Labor

Statistics, 2011), attention of recent research has turned to identifying whether female managers

have an impact on workplace gender inequality. Many studies have argued that female managers

contribute to more equitable outcomes for other women, thus providing a potential lever for

reducing gender inequality (e.g., Ely, 1995; Cotter et al., 1997; Nelson and Bridges, 1999; Hultin

and Szulkin, 1999, 2003; Cohen and Huffman, 2007). These studies infer the impact of female

managers based on observations that a greater proportion of women in management in a given

Abraham

4

industry (e.g. Cohen and Huffman, 2007) or firm (e.g. Hultin and Szulkin, 1999) is associated

with a lower degree of gender inequality in terms of wages (e.g. Hultin and Szulkin, 1999, 2003;

Cohen and Huffman, 2007; Shin, 2012) or job segregation (e.g. Baron et al., 1991; Pfeffer,

Davis-Blake, and Julius, 1995; Kulis, 1997; Huffman, Cohen, and Pearlman, 2010). Drawing on

theories of homophily, in-group bias, and gender, which largely suggest that female managers

will act in a way that leads to advantages for female employees, extant studies have assumed that

female manager influence accounts for this observation that the presence of more women in

management is associated with less gender inequality.

Past studies have proposed two probable pathways through which female managers may

attenuate various forms of gender inequality. First, female managers are posited to have a direct

impact through interactions with, and evaluations of, non-managerial employees. Because of

gender-based homophily (e.g. McPhereson, Smith-Lovin, and Cook, 2001) and in-group

preference (e.g. Tajfel and Turner, 1979), past studies posit that female managers facilitate

interpersonal interactions (Hultin and Szulkin 1999, 2003; Ely 1994) and serve as mentors and

sponsors (Ibarra, 1995; Ragins and Scandura, 1999; Johnson and Scandura, 2004; McGinn and

Milkman, 2012) for female non-managerial employees. Consistent with this perspective, Hultin

and Szulkin (1999) attributes their finding of less gender wage inequality in Swedish

organizations with a greater presence of women among managers partially to the interactional

benefits to women of having other women in high-ranking organizational positions. They state

that "female subordinates should be advantaged when other women are an integral part of the

organization's power structure, simply because interaction within organizations is facilitated by

gender similarity between actors" (Hultin and Szulkin, 1999: 459-60). Similarly, Huffman,

Cohen and Pearlman (2010) posits that there may be less gender segregation in organizations

Abraham

5

where there are more women in management because female managers are more apt to advocate

for female non-managerial employees in those settings. Other studies suggest that while male

employees are typically seen as more competent or valuable organizational members on the basis

of their gender (Ridgeway, 1991, 1997), in-group biases may lead male and female managers to

evaluate employees differently (e.g. Tajfel and Turner, 1979). Consistent with this perspective,

Shin (2012) argues that having women on the board of directors leads to more equitable

executive wages because female board members evaluate other women more favorably than do

male board members.

A second way that female managers are proposed to attenuate gender inequality is indirectly, by

reducing the prevalence of common gender stereotypes and leading to more gender equitable

organizational reward structures. For example, Stainback and Kwon (2012) argues that the

observed association between female managers and gender-based job segregation in their study

is consistent with the argument that female managers reduce gender stereotypes and in-group

preferences that typically advantage men. This perspective is based on existing research that

posits that female managers act to decrease the salience of gender as a category affecting the

views of both managerial and non-managerial employees (Ely, 1995) by reducing gender

stereotyping (Ely, 1994; Blau, Ferber, and Winkler, 2006; Konrad, Kramer, and Erkut, 2008).

Past studies have also attributed the observed association between the female presence in

management and gender inequality to the role of female managers in establishing more equitable

organizational policies and reward structures (Hultin and Szulkin, 1999, 2003). This argument is

based on insights from organizational and sociological research that indicates that members of

top management play a considerable role in setting organizational pay rates (Bridges and Nelson,

1989; Nelson and Bridges, 1999), and that female managers, in particular, are more likely to

Abraham

6

push for gender equity initiatives (Cohen and Huffman, 2007; Dobbin and Kalev, 2009), even

when facing the same organizational and market constraints (Baron et al., 1991).

While most research in this tradition concludes that women in positions of organizational power

serve to attenuate gender inequality, some studies offer conflicting findings (e.g. Shin, 2012;

Stainback and Kwon, 2012; Blau and DeVaro, 2007). But, even these studies with more

pessimistic conclusions about the role of female managers for attenuating gender inequality

attribute findings to manager influence. Some studies highlight that female managers face both

preferences and barriers, which may preclude them from leading to greater gender equity. For

example, while Shin (2012) finds an association between the gender composition of the board of

directors and the executive wage gap, there is no evidence of a difference in the executive wage

gap based on the gender of the CEO. The absence of an association between having a female

CEO and gender inequality in this study is attributed to the unavailability of the CEO for

mentoring. Implicit in this line of reasoning is that we would see a reduction in the wage gap in

settings with female leaders if those women were available to mentor. Similarly, Penner and

colleagues (2012) find that there is no difference in gender wage inequality among supermarket

workers based on the gender of the manager. Drawing on status characteristics theory (e.g.

Berger et al., 1977; Ridgeway and Correll, 2004, 2006; Heilman and Hayes, 2005; Ridgeway et

al., 2009), they argue that female managers may not attenuate wage inequality because, like male

managers, they devalue work performed by female employees and perceive male employees to

be more valuable contributors.

Together, these arguments suggest that any association between the representation of female

managers and the degree of gender inequality is the result of some form of female manager

influence. While it is plausible that managers, and female managers in particular, affect the

Abraham

7

degree of gender inequality, this conclusion has not been directly tested in extant research.

Rather, studies have inferred that the mechanism driving the observed associations between the

presence of women in management and gender equality is the result of female manager influence

and decision making. Without direct evidence that female managers attenuate gender inequality,

it is possible that alternative processes are driving the observed association between women in

management and gender inequality. Specifically, unobserved organizational attributes may

simultaneously impact the degree of gender inequality and the presence of women in

management in a given setting. Developing an accurate theory of the role of female managers for

workplace gender inequality, therefore, requires a direct examination that takes the

organizational setting into account.

ALTERNATIVE MECHANISM: ORGANIZATIONAL PROCESSES

The significance of organizational processes and practices for shaping gender inequality has been

clearly established (e.g., Baron and Beilby, 1980; Bridges and Nelson, 1989; Nelson and

Bridges, 1999; Reskin, 2003; Castilla, 2008; Kalev, 2009; Castilla, 2011). Organizational

resources, including wages and job opportunities, are allocated on the basis of the organization’s

reward structure (e.g. Gibbons, 1998; Castilla, 2008). Scholars have argued that organizational

practices and policies mediate the impact of both micro and macro social structure and processes

(Baron and Beilby, 1980), thus affecting the degree of workplace inequality (e.g. Beilby, 2000;

Reskin, 2000; Reskin and McBrier, 2000). At times the formal rules of bureaucratic workplace

settings have been posited to lessen the effects of individual workers’ characteristics for gender

inequality (e.g. Bielby and Baron, 1986; Huffman and Cohen, 2004; Kmec, 2005), mainly by

limiting employers’ ability to act on their stereotypes and biases (Reskin, 2003). On the other

hand, organizational processes intended to increase diversity and minimize inequality may serve

Abraham

8

to maintain the status quo, such as diversity policies (Kalev, Dobbin, and Kelly 2006), cross-

training programs (Kalev 2009), and merit-based pay practices (Castilla 2008).

Because organizational processes play a central role for gender inequality, any account of the

effect of manager influence on gender inequality failing to take the organizational context into

consideration rests on weak grounds. In the absence of direct evidence, existing accounts are

vulnerable to the possibility that organizational processes, and not manager influence, are driving

the association between the presence of women in management and gender inequality.

Organizations with more women in management may have less gender inequality not because

the female managers are having an impact, but rather because the same organizational processes

driving women into management are also reducing gender inequality among non-managerial

employees. In examining women’s access to management, some studies argue that there are

invisible barriers to entry, commonly referred to as "glass ceilings", barring women from

ascending the organizational hierarchy and obtaining high-status positions (Morrison et al., 1987;

Powell and Butterfield, 1994; Cotter et al., 2001; Albrecht, Björklund, and Vroman, 2003). The

degree to which these barriers preclude women from accessing managerial positions has been

found to depend on various attributes of the specific organizational setting. For example, more

formalized work processes and objective criteria are proposed to minimize inequality in access to

management positions, as stated by Reskin and McBrier (2000: 214): “Insofar as formal

employment practices require employers to standardize procedures and achievement criteria,

managerial sex composition in organizations with highly formalized personnel practices will be

less affected by ascription". Studies in both economics and management posit that the degree to

which the organization is “family-friendly” (Smith, Smith, and Verner, 2013), the presence of

organizational initiatives including affirmative action plans and diversity committees (Kalev,

Abraham

9

Dobbin, and Kelly, 2006), and women’s representation among board members (Matsa and

Miller, 2011; Skaggs, Stainback, and Duncan, 2012) also contribute to facilitating women’s

access to senior management positions.

To the extent that organizational processes either facilitate or hinder women’s access to

managerial positions in a given setting, those same processes may also impact other forms of

workplace gender inequality. The underrepresentation of women in management roles, for

example, has been attributed to recruitment and outreach practices that do not seek a broad range

of potential workers (Federal Glass Ceiling Commission, 1995; Sturm, 2001). The use of

referrals to generate applicant pools for job vacancies provides one such recruitment strategy,

particularly for organizations characterized by gender segregation. Because individuals tend to

have same-sex contacts (McPherson, Smith-Lovin, and Cook, 2001), organizational reliance on

referrals from current employees tends to generate an applicant pool with a gender composition

that resembles that of the organization (Reskin and McBrier, 2000; Marsden and Gorman, 2001;

Torres and Huffman, 2002; Fernandez and Sosa, 2005). Therefore, this recruitment strategy may

simultaneously impact the degree of gender segregation among both managerial and non-

managerial roles. Similarly, the standardization of practices that establish managerial

responsibility have been argued to both increase managerial diversity (Kalev, Dobbin, and Kelly,

2006; Reskin and McBrier, 2000) and to attenuate job segregation and wage inequality (Bielby,

2000; Reskin, 2000, Castilla, 2008, 2010). Given that organizational attributes may not be as

easily observable to researchers examining gender inequality, these organizational features may

be masked by the more readily observable measure of women in management. Cohen and

Huffman (2007) highlight this concern stating (2207: 698): “endogeneity could confound our

Abraham

10

results to the extent that unmeasured variables are driving both the gender wage gap and the

representation of women in management.”

In addition to impacting gender inequality, organizational attributes govern the degree of control,

or power, managers have over allocating resources to employees. Formalized organizational

processes, for example, are associated with not only less gender inequality, but also lower levels

of managerial discretion (Beilby, 2000; Reskin and McBrier, 2000). Because differences in

gender inequality outcomes can only be attributed to manager influence to the extent that

managers have control over their subordinates’ wages, hiring, and promotion, understanding

managerial power is critical to any examination of the impact of female managers on gender

inequality. Power can be broadly defined as having control over other actors, resources, and

things (Wolf and Fligstein, 1979). One critical aspect of managerial power for studies examining

differences in the effect of female, relative to male, managers on workplace gender inequality is

that this power stems from the actor’s position as manager. Weber ([1914] 1968) defined this

type of positional power as “legitimate authority” as it is not innate to the actor, but rather arises

from the structural position that an actor occupies. While generally true, the scope and bounds of

that power is context specific (Cohen and Huffman, 2007). Yet extant research has largely made

the critical assumption that managers have control over the allocation of resource to subordinates

(e.g. Baron et al., 1991; Hultin and Szulkin, 1999; Cohen and Huffman, 2007; Penner, Toro-

Tulla, and Huffman 2012).

Stainback and Kwon’s (2012: 224) multi-organization study provides support for the importance

of knowing the nature of managerial power in a given organizational setting. They find that

having more women among managers, a position typically “granted substantial organizational

power", is associated with less gender inequality, whereas having more women among

Abraham

11

supervisors, who tend to “hold substantially less organizational power", does not have an impact

(Stainback and Kwon, 2012). In the case of Penner and colleagues’ (2012) study using within-

organization employment records for a large U.S.-based grocery retailer, limited managerial

control over wages may provide an alternative explanation for their finding that female managers

do not reduce wage inequality. While this research matches employees to managers, the grocery

retailer they study operates under collective bargaining agreement, which may restrict the level

of control managers have over wages.1 It is plausible that their finding in this grocery retailer is

not because female managers devalue women’s work, but rather the result of managers not

having sufficient control over wages in this setting.

Furthermore, to the extent that managers have the power to affect employee outcomes, they are

most apt to have an impact on those employees reporting to them directly. Therefore, in order to

determine whether male and female managers have a differential effect on gender inequality

among non-managerial employees, it is necessary to compare employees reporting directly to

male and female managers. Despite using aggregate industry-level data in their analyses, Cohen

and Huffman (2007) highlights the importance of considering true reporting structures stating

that they “remain interested in within-organization effects of managerial composition on gender

inequality…this more direct effect remains the most plausible pathway by which female

managers influence gendered outcomes.” Similarly, using organizational-level data, Hultin and

1 Penner et al. (2012) makes the claim that managers have control in their research setting; however the retail grocer they study has several distinct features that cast doubt on the appropriateness of this setting for addressing whether female managers impact wage inequality. First, as the original paper analyzing these data states, “all non-managerial employees were covered by collective bargaining agreements” in this setting (Ransom & Oaxaca 2005: 222). Furthermore, this employer was found guilty in a gender discrimination suit having “to pay several million dollars in ‘back pay’” (Ransom & Oaxaca 2005: 221). Lastly, the within job wage ranges for both non-managerial and managerial employees were narrow, with standard deviations for hourly wages ranging from $0.00 to $1.13. Therefore, while the authors do not find that female managers reduce gender inequality, it is plausible that this finding may be attributed to particular features of the setting. For additional details on this setting, see Ransom and Oaxaca (2005).

Abraham

12

Szulkin (2003) finds that the presence of women among lower-level managers, who are likely

the direct managers of non-managerial employees, had a greater effect on wage inequality than

did women among higher-level managers. Yet most studies examining the relationship between

manger gender and gender-based inequality base findings on analyses of datasets that do not

match individual employees to the male and female managers to whom they report (see Penner,

Toro-Tulla, and Huffman, 2012 for an exception). Rather than looking at differences in gender

inequality for subordinates of male versus female managers, these studies proxy manager-

subordinate reporting relationships using aggregate level data at either the industry- or

organization-level to compare whether a higher proportion of women in management is

associated with less gender inequality among non-managerial employees in a given setting (e.g.

Cohen, Broschak, and Haveman, 1998; Hultin and Szuklin, 1999; Giuliano, Levine, and

Leonard, 2006; Cohen and Huffman, 2007; Kurtulis and Tomaskovic-Devey, 2009).

This indirect approach in extant studies confounds mechanisms of managerial impact with

mechanisms related to selection. This is problematic because the observed relationships might be

spurious in nature such that unobserved organizational factors, and not individual female

managers, are driving the findings of extant studies. In order to accurately attribute the

association between women in management and gender inequality to the impact of female

managers, it is necessary to isolate manager influence from mechanisms related to the

organizational setting. I argue that in order to achieve this, studies must examine differences in

gender inequality for those reporting directly to male and female managers, in settings where

managers have unhindered control over relevant resources. Using a unique research setting

where I am able to link non-managerial employees to their immediate manager and where

managers have explicit control over relevant resources, this study unpacks the causal processes

Abraham

13

by which managers impact workplace gender inequality by isolating managerial influence

mechanisms from organizational processes.

METHODS

Research Setting

This study examines the relationship between manager gender and three distinct forms of

gender-based workplace inequality – wage inequality, job segregation, and the relative use of

flexible work arrangements - for non-managerial employees through a case study of multiple

branches in a globally diversified financial services firm (FinServ). FinServ provided the annual

personnel databases for all employees in its U.S. operations for the 41-month period from

January 1996 through May 1999 (the study period). These databases capture demographic

characteristics, including gender, race, and age, as well as employment specific characteristics,

including tenure, salary, and full-time status. In order to accurately understand whether male and

female managers have differential effects on gender inequality in terms of wages, job allocation,

or the use of flexibility, it is necessary to have data on entire work groups where the reporting

structure between manager and non-managerial workers can be definitively identified. Therefore,

this study focuses on a subset of FinServ, namely the retail branches, where the databases

included complete workgroups with organizational codes linking individual managers to each

non-managerial employee reporting to them. Since managers are most likely to have an impact

on those employees reporting to them directly, these data provide a uniquely well-suited setting

for understanding the impact of female managers on gender inequality. Given that such data are

difficult to access, especially across multiple firms, I exploit the richness of data that results from

Abraham

14

focusing on a single firm (see Sørensen, 2000; Fernandez, Castilla, and Moore, 2000; and

Petersen et al., 2000 for similar approaches).

FinServ’s retail branches are client facing offices that offer a wide range of financial products

and services to retail financial services customers. Among others, these services and products

include mutual fund accounts, mortgages and margin lending accounts. These retail branches

employ individuals in 6 unique positions: five non-managerial and one managerial. Each branch

employs an average of approximately eight non-managerial workers (SD = 2.91) each reporting

to a single branch manager. Using several HR databases, I was able to construct a longitudinal

database for all 1,992 non-managerial employees and 156 managers employed in FinServ’s 120

retail branches across eight U.S. states during the study period. This database is unbalanced, with

each employee having between one and four years of data,2 for a total of 3,888 person years -

357 person years are branch managers and 3,531 person years are non-managerial employees.

Since longitudinal models used in the analysis of wage inequality exclude persons with single

observations, all data analyses employed the same exclusion – a total of 882 people3 - in order to

base comparisons on the same cases. The final analyses are based on 3,006 person years – 2,649

non-managerial and 357 managers - or 1,266 persons – 1,110 non-managerial employees and

156 managers.

In addition to the aforementioned personnel data, this analysis draws on several organizational

documents, including manager training documents and FinServ’s Employee Handbook (herein, 2 Employees in this setting may be present for all years in the study, enter the retail branches after the beginning of the study (i.e. left censored), exit the retail branches before the end of the study window (i.e. right censored), or both enter the retail branches after the beginning of the study window and exit before the end of the study window (i.e. both right and left censored). 3 For 882 individuals, there is a single observation as these people are right, left, or both right and left censored. All persons with 1 year of observations were non-managerial employees, therefore none of the excluded persons were managers. The gender composition of these excluded non-managerial employees mirrors the study population as 69.5% of these people are female.

Abraham

15

Handbook), to illustrate the role of managers in this setting. In a branch manager training session

conducted the year prior to the start of the study window, for example, managers were

encouraged to act in an entrepreneurial capacity as they were instructed to "behave like you are

in business for yourself" and to "hold yourself accountable for outcomes." In terms of

compensation in particular, an organizational goal to "empower managers to make compensation

decisions" and to "institute flexible versus mechanical compensation" was reiterated to managers

during this training session. When addressing the degree of control that managers have over

compensation, the Handbook similarly indicates that the "branch manager owns compensation

decisions." In terms of hiring and promotion, the branch manager "recruits, trains, manages, and

develops a high performance team," "identifies staffing levels", and “plans, approves, and

appropriately paces all training." While HR plays a supporting role to “help supervisors review

and plan employment activities” and encourages managers to “give preference to internals,”

managers have control over hiring and promotion decisions as they are empowered to select an

appropriate candidate for an opening. Similarly in terms of allowing employees to use flexible

work options the Handbook indicates that managers “are encouraged to explore and use flexible

work options” and the manager “has the final decision as to who may participate and may adjust

work schedules at any time.” Whereas existing studies make the critical assumption that

managers have control over the decisions impacting relevant outcomes (e.g. Baron et al., 1991;

Hultin and Szulkin, 1999; Cohen and Huffman, 2007; Penner, Toro-Tulla, and Huffman, 2012),

together these organizational statements provide explicit evidence that the branch managers in

this setting indeed have control over wages, allocation to jobs, and use of flexible work options

for their subordinates.

Abraham

16

By examining differences in outcomes for employees reporting to female versus male managers

within FinServ’s retail branches, I am also able to compare structurally equivalent subunits (i.e.

branches). Each manager in this setting oversees a single comparable branch that employs

individuals in each of the non-managerial positions. These managers operate under the same

organizational guidelines, particularly as related to the outcomes of interest in this study.

Therefore, this analysis exploits within organization variation in these subunits, while controlling

for potentially confounding attributes that may vary across branches. In other words, by focusing

on the retail branches of FinServ, this study is better able to disentangle whether the gender of an

employee’s manager has implications for gender inequality by isolating manager gender and

controlling factors such as organizational policies, manager tenure, and work group size.

Together, these features make the retail branches of FinServ well-suited for identifying whether

female managers have an impact on gender inequality among the non-managerial employees

reporting to them.

Dependent Variables

My analyses use three distinct dependent variables to estimate the degree of gender inequality in

terms of wages, job segregation and the use of flexible work arrangements. Each of these

variables varies annually between 1996 and 1999 and was created using data from FinServ’s

personnel database.

Wages. To measure non-managerial employee wages, I used the natural logarithm of pre-tax

hourly wages. Using hourly wages provides a measure for employee compensation that is

independent of differences in the number of hours worked or fluctuations in work schedule.

Abraham

17

Position. In their 1995 study, Peterson and Morgan argue that the gender wage gap is largely

eradicated when men and women in the same job are compared (see also Kilbourne et al., 1994).

In other words, allocative processes account for a large part of wage inequality (Petersen and

Saporta, 2004; Fernandez and Mors, 2008). Therefore, to fully understand whether female

managers impact gender inequality, it is necessary to examine whether the degree of gender-

based job segregation differs among employees reporting to female versus male managers. While

supply-side factors have been found to contribute to the gender composition of jobs (Fernandez

and Sosa, 2005; Fernandez and Abraham, 2010; Barbulescu and Bidwell, 2012), many economic

and organizational sociologists have argued that organizational factors also play a role (Baron,

1984; Bielby and Baron, 1986; Nelson and Bridges, 1999; Tomaskovic-Devey and Skaggs,

1999; Sørensen, 2007). Given that the 5 non-managerial branch positions follow a clear

hierarchy (see Figure 1, positions and the branch hierarchy discussed in detail below), position

was coded as a five-category ordinal variable. Position is coded as follows: 1 for employees in

the lowest branch position of teller, 2 for representatives, 3 for officers, 4 for executives, and 5

for relationship managers.

Part-time. In addition to wages and allocation to jobs, organizational employees may face

gender-based inequities in terms of other resources such as flexible work arrangements. While

one argument against perceiving access to flexibility as a resource is that flexible work is

sometimes linked to penalties for employees,4 there is also evidence that employees working

under flexible work arrangements “earn wages at least equal to their fixed schedule counterparts”

4 In the research setting presented herein, a comparison of hourly wages of part-time and full-time employees reveals that employees to not suffer a wage penalty for using flexible work arrangements. Also, part-time employees working at least 25 hours per week are eligible for all of the benefits available to full-time employees including, medical benefits, company match to retirement savings plan contributions, tuition reimbursement and paid time off.

Abraham

18

(Weeden, 2005). Furthermore, studies have argued that both male and female workers cite

flexibility as the most desirable feature to have in a job (Glass and Estes, 1997; Golden, 2001) as

it enables employees to manage their work and non-work responsibilities. While this preference

exists, opportunities for flexible work vary greatly, both across and within organizations (Kelly

and Kalev, 2006; Wharton, Chivers, and Blair-Loy, 2008). As Galinsky and colleagues (1996)

argue, managerial support is not only important, but may actually be more significant than

formal policies in determining the use of flexibility in organizations. While there has not been

any examination looking at differences in the use of flexibility based on manager gender, some

studies suggest that a difference may exist. For example, some studies suggest that the presence

of formal policies will be higher in organizations that have more women in management (Dreher,

2003). Similarly, Dobbin and Kalev (2009) find that organizations with a higher proportion of

women, particularly white women, in management are more likely to adopt diversity programs.

In this study, flexible work is measured as whether non-managerial employees are employed as

part-time employees. This variable is measured by a dichotomous dummy variable where part-

time equals 1 if the employee has a part-time work arrangement and 0 otherwise. In this setting,

part-time work provides evidence of both reduced work hours and flexibility in scheduling given

that branches operate during normal business hours (i.e. approximately 8:30am-6:00pm).

Therefore, part-time employees work different shifts within this window.

Explanatory and Control Variables

The main explanatory variables are Employee female and Manager female to capture the gender

of the focal employee and the employee’s manager, both coded 1 for female and 0 for male. Both

Employee female and Manager female are used in all models predicting each of the dependent

Abraham

19

variables. The main focus of this study is to determine whether male and female managers

impact outcomes for male versus female employees differently. Therefore, the interaction

between Manager female and Employee female is included to capture this difference.

Controls for wage models. All of the wage models include controls for a number of

characteristics of individual non-managerial employees including a set of dummy variables for

three of the four racial categories: African-American, Asian, Caucasian, Hispanic. The largest

category, Caucasian, is omitted as the reference category. Employee age and Employee tenure

are measured in years for each year from 1996 to 1999 based on the employee’s date of birth and

date of hire, respectively. Marital status is coded 1 if the employee is married in a given year and

0 otherwise. Additionally, the interaction of Employee female and Marital status is included in

all wage models as some gender scholars have found that differences in pay and promotion are

related to expectations about workers family responsibilities (Roth, 2008). The notion of the

“unencumbered ideal worker” (Bailyn, Drago, and Kochan, 2011) suggests that managers may

prefer single male employees most. By including the main and interaction effects for Marital

status, these models account for this possible association.

Models also include a control for Manager tenure, as male and female managers differ in

average tenure. Manager tenure is measured in years, for each year from 1996 to 1999, based on

the manager's date of hire. A third set of controls are related to characteristics of the individual

branches. One argument may be that larger branches are of higher status indicating that

managers of larger branches may differ in some substantive way from managers of smaller

branches. Therefore, all models control for Branch size, coded as number of employees including

manager, for each year from 1996 to 1999. Lastly, all models include state fixed effects for the

state where the branch is located to account for any regional differences in wages.

Abraham

20

Model Specification

I estimate three sets of models to examine whether gender inequality outcomes differ based on

manager gender. First, I use multivariate analyses with the logarithm of pre-tax hourly wages as

the dependent variable to examine whether relative wages for male and female employees differs

by manager gender. Because employees in this setting are nested within branches, individuals

within branches may have more in common than individuals across branches. This indicates that

employees within branches may not be independent, violating a key assumption of ordinary least

squares (OLS), therefore OLS may result in biased estimates of standard errors (Osborne, 2000).

Furthermore, the data I am analyzing are pooled, cross-sectional time series (yearly) data. A

common approach for analyzing data that are structured in this way is to use fixed-effects

models, which would capture within-individual and within-branch, over time variation. Given

my research question, I am inherently interested in the gender of non-managerial employees and

the gender of branch manager, two attributes that are time-invariant.5 Therefore, a fixed-effects

model is not suitable for this study as both employee gender and branch manager gender would

be dropped from such models. Therefore, I estimate various cross-sectional time-series linear

models using the method of generalized estimating equations (GEE).6

5 Theoretically, branch manager may vary in cases where a branch is managed by a female manager in one year and a male manager in a subsequent year (or vice versa), for example. During this time period, however, these switches in manager gender were very rare. Only 3 branches, for a total of 4 percent of persons, experienced such an event. Furthermore, the direction of this switch in branch manager gender was not consistent: 1 branch went from having a female to a male manager and the other 2 branches went from a male to a female manager. In most cases this switch in manager occurred in either the first or the last year of the study period making the period of time that either the initial or subsequent manager was in charge of the branch insufficient for drawing comparisons between managers within a branch. Therefore, branch fixed effects would not allow for estimating parameters for the effect of manager gender for most cases in this study. 6 Results are largely consistent when estimating models using less structure on the variance-covariance matrix (i.e. OLS).

Abraham

21

As aforementioned, the data that I am analyzing are unbalanced as individuals may enter the

dataset after the start of the study window and/or exit before the close of the study window.

Given that these data include repeat observations for individuals7 over time, I ran diagnostics to

test the assumptions of homoscedasticity and that errors are not auto-correlated across time

periods. As expected, errors in these data are heteroskedastistic, as the variance of the errors

varies across individual non-managerial employees. Additionally, the Woolridge test for

autocorrelation in panel data reveals that there exists first-order auto-correlation, where the errors

at time t are correlated with the errors at time t-1. Thus, longitudinal GEE models accounting for

both heteroskedasticity and first-order autocorrelation are used to estimate the effect of manager

gender on gender-based wage inequality.

Second, I use log-linear hierarchical models to examine whether job segregation or the relative

use of flexible work varies for employees based on manager gender. Log-linear models are a

special class of models that appropriately deal with comparisons of categorical data (Bishop,

Fienberg, and Holland, 1975; Haberman, 1978). Specifically, these models uncover the potential

relationships among categorical variables in a multiway contingency table. In terms of job

segregation, log linear models allow for testing the null hypothesis that manager gender is

independent of the joint distribution of male and female employees to the five non-managerial

branch jobs. In order to test whether the gender composition of jobs across the branch hierarchy

is the same for employees reporting to female and male managers, I estimate the following log-

linear model comparing the relationship between Employee female and Branch position by

Manager female

7 For 882 individuals, there is a single observation as these people are right, left, or both right and left censored. Therefore, when clustering standard errors to account for the fact that errors are correlated across time periods by individual, these 882 individuals are dropped from the analysis. All models, then, are based on the remaining 2,649 person-years for those individuals with 2 or more observation.

Abraham

22

{FB} {M}

where F is Employee female, B is Branch position, and M is Manager female.

Similarly for use of flexible work arrangements, log-linear models allow for testing the null

hypothesis that manager gender is independent of the joint distribution of male and female

employees to part-time work arrangements. In order to test whether the relative use of part-time

by female versus male employees is the same for employees reporting to female and male

managers, I estimate the following log-linear model

{FP} {M}

where F is Employee female, P is Part-time, and M is Manager female.

RESULTS

The branch employees in this setting span 6 distinct organizational positions: teller,

representative, officer, account executive, relationship manager, and branch manager. The first

five positions listed are non-managerial positions, each of which may be either exempt or non-

exempt, and the last position, branch manager, is exempt. Individuals in these non-managerial

positions report directly to the branch manager. Figure 1 depicts the hierarchy of positions, as

claimed in FinServ documents as the career path within a branch, as well as the key

responsibilities of each role. FinServ terms non-managerial positions within the branch "manage

self" positions, as these positions not have any personnel management responsibilities. The

position of branch manager is the only "manage others" position, as a key set of responsibilities

is related to personnel management and managing the branch team.

[Insert FIGURE 1 about here]

Abraham

23

While the positions to the left of the dotted line in figure 1 do not have managerial

responsibilities, there is a clear hierarchy among the non-managerial branch positions in terms of

responsibilities and wages. There is a structure of cumulative skills and knowledge within the

branches of FinServ such that incumbents of each position were expected to be expert in

particular skills associated with their position as well as the skills of each lower level position.

For example, representatives are focused on unsecured products and sales and service, but they

are also responsible for the tasks of tellers, namely basic products and transactions. Table 18

reports the distribution of employees across branch positions and mean hourly wages for each

branch position. Consistent with this branch hierarchy, the mean hourly wage increases steadily

with the lowest position being teller (mean = $12.24, SD = $1.82) and the highest non-

managerial position being relationship manager (mean = $32.17, SD = $7.90). The mean hourly

wage increases by a minimum of $2.65 (mean of $4.35) for each increase in non-managerial

position along this hierarchy. In terms of the distribution of non-managerial employees across the

branch hierarchy, table 1 shows that over 70 percent of all employees occupy the lower-level

positions of teller, representative and officer.

[Insert TABLE 1 about here]

The branch managers comprise approximately 15 percent of all employees in the retail branches.

Of particular interest for this study, each branch is head by a single manager and these managers

are the sole employees in the retail branches with personnel responsibilities overseeing all other

employees and branch operations. Table 2 reports descriptive statistics for branch managers, both

overall and by manager gender. The average hourly wage for managers was approximately

8 For simplicity, all descriptive statistics are presented for 1999. The patterns presented in each of the descriptive tables for 1999 are consistent with similar tables constructed for each of the 4 years, 1996 – 1999, included in this study.

Abraham

24

$33.98 in 1999 (SD = $7.55) with approximately 46 percent of managers being female. Given

that the focus of this study is to determine whether the degree of gender inequality among

employees differs based on the gender of their manager, it is important to identify if male and

female managers in this setting differ in any substantive ways other than gender. While female

managers earn slightly less than male managers on average, this difference in wages is not

significant. Similarly, male and female managers do not differ significantly in either their racial

composition, with both groups being equally diverse in terms of race, or their mean age. Female

managers are more tenured, however, having over four years more tenure (P<.05), on average,

when compared to male managers. To the extent that more experienced managers have more

power within the organization, this difference indicates that female managers should possess at

least as much control over resource allocation as male managers.

[Insert TABLE 2 about here]

The first column of table 3 presents overall descriptive statistics for male and female non-

managerial employees irrespective of whether they report to a male or female manager. The

second and third columns of table 3 separate non-managerial employees based on the gender of

their manager to determine whether the employees reporting to male and female managers differ

in any substantive way.9 As column 1 of table 3 reveals, on average there are eight non-

managerial employees in each branch and approximately 70 percent of all non-managerial

employees are female. Comparing overall mean hourly wages across all jobs for female and male

employees indicates that female employees earn approximately 74 cents for every dollar that

men earn (P<.001). Female employees are nearly five times as likely to work part time (P<.05),

9 Educational information (i.e. highest degree) is only populated for 35 percent of employees included in the study population. This information is provided by employees on a voluntarily basis. Since it is unclear whether this information is missing at random, education is not included in the main analyses.

Abraham

25

are about two years older, and have nearly two more years of tenure on average (P<.05) than

men. Male and female employees do not differ, however, in their likelihood of being married or

in terms of racial composition.

[Insert TABLE 3 about here]

Given the aim of this study, it is necessary to not only determine whether male and female non-

managerial employees differ, but also whether male and female managers oversee comparable

individuals. The next two columns of table 3 present key descriptive statistics for all variables of

interest for employees by branch manager gender allowing for a comparison of female non-

managerial employees, as well as male employees, reporting to female versus male managers.

The final column of table 3 presents the results of t-tests of these within non-managerial

employee gender comparisons. Male and female managers oversee branches that are similarly

female, but female managers manage smaller branches on average (7 versus 8 employees,

P<.001). Neither male nor female non-managerial employees reporting to female managers differ

significantly in terms of wages, the propensity to work part-time, or race when compared to

those reporting to male managers. Female non-managerial employees reporting to female

managers do have longer tenure on average (8 vs. 6.2 years, P<.05) than female employees

reporting to male managers. While average tenure for male employees does not differ, male

employees reporting to female managers are older (38.6 vs. 35.2 years, P<.05) and more likely to

be married (57.8 vs. 39.5 percent, P<.05) than those men reporting to male managers.

Female Managers and Wage Inequality

Table 4 presents results from the three models estimating the effect of manager gender on non-

managerial employee log hourly wages. Model 1 reveals that, irrespective of manager gender,

Abraham

26

non-managerial female employees earn 14.9 percent (p < .001) less than their male counterparts.

In order to assess whether female managers impact this evident gender inequality in wages,

Model 2 introduces both the main effect of manager gender and the interaction effect between

manager gender and gender of the non-managerial employee. First, the negative effect of female

remains, indicating that among employees reporting to male managers (the reference category

for manager female), female employees earn approximately 16 percent (p < .05) less than male

employees. The negative main effect of manager gender indicates that non-managerial male

employees (the reference category for female) reporting to female managers earn slightly over

five percent less (p < .05) than those reporting to male managers.

[Insert TABLE 4 about here]

The relevant comparison for testing the proposition that female managers will reduce wage

inequality, however, is whether the wage gap between male and female employees is lower

among those reporting to female managers as compared to those reporting to male managers.

Given that the coefficient of the interaction term is insignificant, we can conclude that this is not

the case. While employees reporting to female managers earn less than those reporting to male

managers, the relative wage for female versus male employees, or the gender wage gap, does not

differ based on manager gender.

Thus far, this analysis has not compared male and female non-managerial employees working in

the same organizational positions, or jobs. As aforementioned, the gender wage gap has been

shown to virtually disappear when men and women in the same job are compared (e.g. Kilbourne

et al., 1994; Petersen and Morgan, 1995; Petersen and Saporta, 2004). Therefore, to fully

understand whether female managers impact wage inequality, it is necessary to compare wages

Abraham

27

for male and female non-managerial employees within branch position. Model 3 is a fully

interacted model estimating the effect of manager gender on male versus female employee

wages, introducing dummy variables for the non-managerial branch positions (teller is the

omitted category). It is worth noting that the positive relationship between each of the dummy

variables for the non-managerial positions indicates that male representatives, officers, account

executives, and relationship managers reporting to male managers earn 16, 37, 58, and 79

percent more, respectively, than male tellers reporting to male managers. This pattern is in line

with the pattern revealed in Table 1, that mean wages for each job are significantly higher than

mean wages for the reference category of teller. Since teller is the omitted branch position, the

observed negative relationship between employee gender and wages reflects the within-job

gender wage gap for tellers reporting to male managers (the reference group for manager

female). Among tellers reporting to male managers, female tellers earn five percent less (p < .05)

than male tellers. The within-job gender wage gap for each of the other four branch positions is

presented by the two-way interactions between employee gender (female) and each of the branch

position. These results indicate that other than tellers, gender wage inequality only exists among

executives with female executives earning nearly eight percent less than their male counterparts.

Therefore, consistent with extant theory, once we compare wages for men and women working

in the same branch positions, wage inequality is attenuated.

In order to determine whether female managers reduce wage inequality within branch position, it

is necessary to compare the relative wages of male and female employees reporting to female

versus male managers. These effects are represented by the coefficients of manager gender (main

effect, two-way interactions, and three-way interactions). The negative main effect of manager

gender indicates that tellers reporting to female managers earn three percent less (p < .10) than

Abraham

28

those reporting to male managers. Additionally, the positive coefficient of the interaction of

manager gender and employee gender indicates that female tellers earn nearly four percent more

when reporting to a female versus a male manager. Together these results indicate that female

managers attenuate gender wage inequality among tellers. The only other position where there is

evidence of gender wage inequality is among executives, with female executives earning nearly

eight percent less than male executives. This wage inequality, however, does not differ based on

manager gender.

To illustrate how this gender wage inequality among tellers differs based on whether employees

report to a female or a male manager I used the coefficients from model 3 (Table 4) to construct

the two-by-two table presented in table 5. This table compares the mean hourly wages for the

typical10 male and female teller based on whether they report to a male or female manager. These

figures allow for a comparison of the gender wage gap among tellers reporting to female relative

to male managers. Among tellers reporting to female managers, female tellers earn 99 cents for

every dollar that male tellers earn. Comparatively, female tellers earn 95 cents for every dollar

that male tellers earn when employees reporting to male managers are considered. This result

provides some support that female managers attenuate wage inequality for subordinates when

compared to their male counterparts. There is less gender inequality in terms of wages for

subordinates reporting directly to female managers, however, this effect is limited to the lowest

position in the organizational hierarchy, that of teller.

Female Managers and Job Segregation

10 Predicted values for hourly salary were calculated using Model 3 (Table 5) for a typical teller profile: a white, single individual, working fulltime, with average tenure and age reporting to a manager with average tenure. The resulting natural logarithm of hourly wages for each group (e.g. men reporting to male managers) was then exponentiated for ease of interpretation. This approach, as opposed to simply calculating means for groups, allows mean hourly wages to take into account all controls from Model 3.

Abraham

29

In terms of the distribution of female non-managerial employees across the levels of the branch

hierarchy, figure 2 illustrates the degree to which jobs are segregated in terms of gender, by

branch position. Consistent with the glass ceiling hypothesis, the proportion of women in lower-

level positions of teller and representative is considerably higher than the proportion of women

in the higher-level positions of account executive and relationship manager (p<.01; LR chi-

square = 163.35, d.f. = 4). While this pattern is evident, this is simply a descriptive pattern as

these are post-hire data which do not provide insight into the mechanisms leading to this

apparent glass ceiling (for a similar discussion of the limitations of post-hire data see Fernandez

and Weinberg, 1997; Fernandez and Abraham, 2010, 2011). For example, not knowing the

proportion of women that are in the consideration set for higher-level positions (i.e. relationship

manager) makes it equally plausible that 1) women are simply not applying to higher level jobs

(as has been shown by Barbulescu and Bidwell, 2012 and Fernandez and Abraham, 2010, 2011,

for example) or 2) managers are showing a preference in selection for male applicants for higher

level jobs. In order to isolate the mechanisms at play in creating these observed patterns it is

necessary to examine pre-hire data (see Fernandez and Abraham, 2010, 2011 and Fernandez and

Weinberg, 1997 for an example).

[Insert FIG 2 and FIG 3 about here]

In the absence of data on the applicant pools, however, these analyses can at least determine

whether the observed glass ceiling pattern differs based on whether employees report to a male

versus a female manager. As figure 2 illustrates, this descriptive pattern appears to be quite

similar for subordinates of both female and male managers. Model 1 in table 6 presents the

likelihood ratio chi-squared statistics and related p-values for the log-linear independence models

for each of the four years included in these data to test whether the gender composition of jobs

Abraham

30

differ based on manager gender. This model tests the null hypotheses that the distribution of

male and female employees across the five levels of the branch hierarchy is independent of

manager gender. While I am not able to draw any conclusions regarding the impact of female

relative to male managers on this distribution of women across levels of the organizational

hierarchy, the insignificant likelihood ratio chi-square for each year indicates that this

independence model is a good fit for the data. The distribution of employees across the hierarchy

does not differ based on the gender of the manager, therefore, to the extent that non-managerial

jobs are skewed in terms of gender, they are equally segregated regardless of the gender of the

branch manager.

[Insert Table 6 about here]

Female Managers and Flexible Work Arrangements

Figure 4 shows the proportion of non-managerial employees, both male and female, working

part-time by manager gender. Employees are nearly twice as likely to work part-time (5.9

percent vs. 10.8 percent) when reporting to a female versus a male manager (p<.1; LR chi-square

= 3.17, d.f. = 1). Figure 5 shows the relative use of part-time by male and female non-managerial

employees reporting to female versus male managers. The proportion of female, relative to male,

employees working part-time is significantly higher among employees reporting to both female

(p<.1; LR chi-square = 3.01, d.f. = 1) and male managers (p<.05; LR chi-square = 5.33, d.f. = 1)

indicating that women are more likely to work part-time. Unsurprisingly, these differences in the

use of flexibility suggest that there is an association between gender and flexibility. They do not,

however, indicate whether female managers are more egalitarian in allowing their male and

female subordinates to use flexible work arrangements. As with the job segregation analyses, this

Abraham

31

is simply a descriptive pattern as these data do not provide insight into the mechanisms leading

to this outcome. For example, not knowing the proportion of men versus women that requested

part-time work makes it equally plausible that 1) men are requesting part-time work at lower

rates or 2) managers are showing a preference for female employees in deciding who to allow to

work part-time. In order to isolate the mechanisms at play in creating these observed patterns it

would be necessary to identify which employees requested flexible work arrangements (for a

similar discussion of the limitations of post-hire data see Fernandez and Weinberg 1997;

Fernandez and Abraham 2010, 2011).

[Insert FIG 4 and FIG 5 about here]

In order to assess whether this evident gender inequality in use of flexibility differs for those

reporting to female versus male managers, Model 2 in table 6 presents the likelihood ratio chi-

squared statistics and related p-values for the log-linear independence models for each of the four

years included in these data. This model tests the null hypotheses that the distribution of male

and female employees to part-time work is independent of manager gender. While I am not able

to draw any conclusions regarding the impact of female relative to male managers on this

distribution of men and women to part-time work, the results indicate that the relative proportion

of men and women working part-time does not differ based on the gender of the manager. In

other words, to the extent that there exists gender inequality in the use of flexibility in this

setting, this inequality is consistent for employees reporting to both male and female managers.

DISCUSSION

Given the rise in women’s presence among managers, this study aims to uncover whether, and

under what conditions, female managers reduce gender inequality among the non-managerial

Abraham

32

employees reporting to them. While recent studies have attributed the lower levels of gender

inequality in settings where there are more women in management to the impact of female

managers, these studies have not provided direct evidence for this mechanism. Using unique

personnel data from 120 branches of a large retail financial services firm where the reporting

structure is identified and managers have authority over employee outcomes, I find that female

managers contribute to differences in outcomes for subordinates in two distinct ways. First, they

provide greater access to equitable wages, but only among employees in the lowest level

organizational position. While the overall gender wage gap does not differ based on manager

gender, female managers do attenuate gender inequality in wages among tellers. Second, they

provide greater access to flexible work arrangements for both male and female employees. While

I do not find that female managers are more equitable in term of access to flexible work

arrangements, both male and female employees reporting to female managers are nearly twice as

likely to work part-time.

These findings have several implications for research on the organizational bases of gender

inequality and workplace flexibility. This case-study design provides the first direct evidence of

the relationship between the gender of the manager and the gendered outcomes of subordinates

in a setting where actual manager-subordinate relationships are identified and managers have

control over outcomes. By illustrating the power of an organizational grounded approach for

developing a more complete understanding of the role of women in management on gender

inequality of subordinates, this study serves as a useful corrective to past approaches which seek

to address the organizational bases of gender inequality employing only indirect evidence on

these questions.

Abraham

33

In terms of the impact of female managers on wage inequality, this study ascertains the

importance of identifying managerial control over setting wages. Given Penner and colleagues’

(2012) finding that female managers have no impact on wage inequality among the low-skill,

supermarket workforce they study, my finding that female managers attenuate wage inequality

among tellers is somewhat surprising. This study casts doubt on Penner et al.’s (2012) claim

that, despite the collective bargaining agreement, female managers in their setting have the

necessary control to impact wage inequality. Given that differences in wage inequality can only

be attributed to the managers to the extent that those managers have control over allocating

wages, their finding may result from the fact that managers lack control over wages in this

setting.

Furthermore, attempting to identify whether female managers attenuate gender wage inequality

without looking at within-job wage inequality may lead to an incomplete answer. If conclusions

about the impact of female managers were drawn solely by comparing the overall wage

inequality across all branch positions in this setting, one would conclude that female managers

do not reduce gender inequality. A more careful examination of wage inequality within specific

branch positions, however, reveals boundary conditions regarding when female managers may

attenuate inequality, namely among employees in lower level organizational positions. Future

research identifying additional boundary conditions for when female managers are more apt to

impact gender inequality is needed.

While these data do not allow for identifying the specific mechanisms of manager influence,

previous research is informative in identifying a potential explanation for the finding that female

managers only attenuate wage inequality in the lower organizational ranks. Value threat theory

suggests that as the lower status group, women may fear that others will not perceive them as

Abraham

34

valuable members of the organization making them less apt to support other women within the

organization (Reagans, 2005; Duguid, 2011; Duguid, Lloyd, and Tolbert 2012). Therefore,

women in positions with the power to affect outcomes of subordinates may act similarly to their

male counterparts, even if they do not hold gender stereotypes disadvantaging women. It is

plausible that female managers feel less threat breaking with the norm of wage inequality among

employees in the lower-level branch position of teller than in higher-level positions, such as

account executive. While this is plausible, additional research is necessary to identify whether

female managers decisions to reduce inequality is driven by concerns over their own

organizational value. Interviews with both male and female managers would be informative for

identifying the degree to which female managers are more concerned with how they are

perceived by other organizational actors. A second possibility is to compare the impact of female

managers on gender inequality across settings where their status relative to male counterparts

varies.

By examining gender differences in the allocation of an understudied resource, namely

flexibility, by managers, this study also broadens our knowledge of whether female managers

impact resource allocation to subordinates differently from male managers. This finding, at a

minimum, suggests that employees reporting to female managers may be more able to leverage

flexible work options. As aforementioned, without knowledge of how employees were given

flexibility it is unclear whether employees reporting to female managers requested flexibility at a

higher rate or if female managers are more accommodating in extending flexibility. However,

irrespective of which mechanism is leading to the higher rate of flexibility among those reporting

to female managers, this finding indicates that working for a female manager increases an

employee’s likelihood of using a flexible work arrangement. In order to isolate the mechanisms

Abraham

35

at play, future research should aim to examine differences not only in flexibility use but in

requesting flexibility for those reporting to female as compared to male managers.

Additionally, the higher likelihood for employees reporting to female managers to use flexibility

may be particularly beneficial for women. In recent studies, women have been found to

experience depletion, or detrimental effects from their multiple roles, whereas men tend to

experience enrichment, or positive spillover effects (Rothbard, 2001). Therefore, the fact that

employees use flexibility more when reporting to a female manager may serve as an indirect

benefit for female employees reporting to female, as opposed to male, managers. To the extent

that flexibility is more valued by women as it reduces the negative experience from balancing

work and non-work responsibilities, the higher likelihood for employees to use flexibility when

reporting to a female manager may be more beneficial for women, than for men, reporting to

female managers.

In addition to advantages to the individual employee, some studies have posited that flexibility is

associated with positive organizational outcomes. For individuals who prefer greater

segmentation between their work and non-work lives, flexible work options that facilitate this

separation have been found to lead to greater organizational commitment (Rothbard, Phillips,

and Dumas, 2005). This suggests that if female managers are both extending flexibility more and

to the employees most in need of segmentation, they may positively impact both their

subordinates and organizational outcomes. There is promise for future research to unpack the

managerial process of allocating flexibility to male and female employees. A more in-depth

examination of this process in a setting with the same features of the organization studied here,

namely where reporting structures are identified and managers have control over allocating

flexibility, is necessary. Specifically, gaining access to formal flexible work arrangements would

Abraham

36

allow for a comparison of different forms of flexibility and provide insight into employee, as

well as employer, preferences in terms of flexible work arrangements.

Abraham

37

References

Albrecht, J., A. Björklund, S. Vroman. 2003 “Is There a Glass Ceiling in Sweden?” Journal of Labor Economics, 21: 145-177. Bailyn, L., R. Drago, and T. Kochan. 2011 "Integrating Work and Family Life: A Holistic Approach." Report for Alfred P. Sloan Foundation Work-Family Policy Network. Barbulescu, R., and M. Bidwell. 2013 “Do Women Choose Different Jobs from Men? Mechanisms of Application Segregation

in the Market for Managerial Workers.” Organization Science, 24:737-756. Baron, J. N. 1984 “Organizational Perspectives on Stratification.” Annual Review of Sociology, 10:37-69. Baron, J. N., and W. T. Bielby. 1980 “Bringing the Firms Back in: Stratification, Segmentation, and the Organization of Work.” American Sociological Review 45:737-765. Baron, J. N., B. S. Mittman, and A. E. Newman. 1991 “Targets of Opportunity: Organizational and Environmental Determinants of Gender Integration within the California Civil Service, 1979-1985.” The American Journal of Sociology, 96:1362-1401. Berger, J. M., H. Fisek, R. Z . Norman, and M. Zelditch, Jr. 1977 Status Characteristics and Social Interaction. New York: Elsevier. Bielby, W. T. 2000 “Minimizing Workplace Gender and Racial Bias." Contemporary Sociology, 29:120-129. Bielby, W. T., and J. N. Baron. 1986 “Men and Women at Work: Sex Segregation and Statistical Discrimination.” The American Journal of Sociology, 91:759-799. Bishop, Y.M. M., S.E. Fienberg, and P.W. Holland. 1975 Discrete Multivariate Analysis: Theory and Practice. M.I.T. Press, Cambridge, MA. Bjerk, D.

Abraham

38

2008 “Glass ceiling or sticky floors? Statistical discrimination in a dynamic model of hiring and promotion.” Economic Journal, 118(530): 961–82. Blau, F. D., and J. DeVaro. 2007 “New Evidence on Gender Differences in Promotion Rates: An Empirical Analysis of a Sample of New Hires.” Industrial Relations, 46(3):511-550. Blau, F. D., M. A. Ferber, and A. E. Winkler. 2009 The Economics of Women, Men, and Work. 6th ed. Prentice Hall. Bridges, W. P., and R. L. Nelson. 1989 “Markets in Hierarchies: Organizational and Market Influences on Gender Inequality in a State Pay System.” The American Journal of Sociology 95(3): 616–658. Briscoe, F., and K. Kellogg. 2011 "The Initial Assignment Effect: Local Employer Practices and Positive Career Outcomes for Work-Family Program Users." American Sociological Review, 76: 291-319. Bureau of Labor Statistics. 2011 Women in the Labor Force: A Databook. www.bls.gov/cps/wlf-databook-2011.pdf Cardoso, A. R., and R. Winter-Ebmer. 2007 “Female-Led Firms and Gender Wage Policies.” ILR Review, 64 (1). Castilla, E. J. 2008 “Gender, Race, and Meritocracy in Organizational Careers.” The American Journal of Sociology, 113 (6): 1479–1526. Castilla, E. J. 2011 “Bringing Managers Back In.” American Sociological Review, 76 (5): 667 –694. Cohen, L. E., J. P. Broschak, and H. A. Haveman. 1998 “And Then There were More? The Effect of Organizational Sex Composition on the Hiring and Promotion of Managers.” American Sociological Review, 63:711-727. Cohen, P. N., and M. L. Huffman. 2007 “Working for the Woman? Female Managers and the Gender Wage Gap.” American Sociological Review 72:681-704. Cotter, D. A., J. DeFiore, J. M. Hermsen, B. M. Kowalewski, and R. Vanneman.

Abraham

39

1997 “All Women Benefit: The Macro-Level Effect of Occupational Integration on Gender Earnings Equality.” American Sociological Review, 62:714-734. Dreher, G. 2003 “Breaking the glass ceiling: The effects of sex ratios and work-life programs on female leadership at the top.” Human Relations, 56:541-562. Duguid, M. M. 2011 “Female Tokens in High-prestige Work Groups: Catalysts or Inhibitors of Group Diversification?” Organizational Behavior and Human Decision Processes, 116 (1): 104–115. Duguid, M. M., D. L. Loyd, and P. S. Tolbert. 2012 “The Impact of Categorical Status, Numeric Representation, and Work Group Prestige on Preference for Demographically Similar Others: A Value Threat Approach.” Organization Science, 23 (2): 386–401. Equal Employment Opportunity Commission. 2004 “Glass Ceilings: Status of Women as Officials and Managers in the Private Sector” Ely, R. J. 1994 “The Effects of Organizational Demographics and Social Identity on Relationships Among Professional Women.” Administrative Science Quarterly, 39: 203–238. Ely, R. J. 1995 “The Power in Demography: Women's Social Constructions of Gender Identity at Work.” The Academy of Management Journal, 38: 589-634. Federal Glass Ceiling Commission. 1995 Good for business: Making full use of the nation’s human capital, the environmental scan. Washington, DC: U.S. Government Printing Office. Fernandez, R. M., and N. Weinberg. 1997 “Sifting and Sorting: Personal Contacts and Hiring in a Retail Bank.” American Sociological Review, 62: 883–902. Fernandez, R. M., and M. B. Abraham. 2010 “From Metaphors to Mechanisms: Gender Sorting In(to) an Organizational Hierarchy.” Paper presented at the annual meetings of the American Sociological Association in Atlanta, Georgia

Abraham

40

Fernandez, R. M., and M. B. Abraham. 2011 "Glass Ceilings and Glass Doors? Internal and External Hiring in an Organizational Hierarchy." Paper presented at the 2011 annual meetings of the American Sociological Association in Las Vegas, NV Fernandez, R. M., and I. Fernandez-Mateo. 2004 “Networks, Race, and Hiring.” American Sociological Review, 71: 42–71. Fernandez, R. M., and M. L. Mors. 2008 “Competing for jobs: Labor queues and gender sorting in the hiring process.” Social Science Research, 37: 1061-1080. Fernandez, R. M., and M. L. Sosa. 2005 “Gendering the Job: Networks and Recruitment at a Call Center.” The American Journal of Sociology 111: 859–904. Fernandez, R. M., E. J. Castilla, and P. Moore. 2000 “Social Capital at Work: Networks and Employment at a Phone Center.” American Journal of Sociology, 105: 1288–1356. Galinsky, E., J. Bond, and D. Friedman. 1996 “The role of employers in addressing the needs of employed parents.” The Journal of social issues, 52:111-136. Gibbons, R. 1998 “Incentives in Organizations.” Journal of Economic Perspectives, 12: 115-132. Giuliano, L., D. I. Levine, and J. Leonard. 2006 "Manager Race and the Race of New Hires." Working Paper. University of Miami, Department of Economics. Glass, J. L., and S. B. Estes. 1997 “The Family Responsive Workplace.” Annual Review of Sociology, 23:289. Golden, L. 2001 “Flexible work schedules: what are we trading off to get them?” Monthly Labor Review,

124:50. Helgeson, S. 1990 The female advantage: women’s ways of leadership. New York, NY: Doubleday.

Abraham

41

Huffman, M. L., and P. N. Cohen. 2004 “Occupational segregation and the gender gap in workplace authority: National versus local labor markets.” Sociological Forum, 19: 121-147. Huffman, M. L., P. N. Cohen, and J. Pearlman. 2010 “Engendering Change: Organizational Dynamics and Workplace Gender Desegregation, 1975–2005.” Administrative Science Quarterly, 55: 255–277. Hultin, M., and R. Szulkin. 2003 “Mechanisms of Inequality: Unequal Access to Organizational Power and the Gender Wage Gap.” European Sociological Review, 19:143–59. Hultin, M., and R. Szulkin. 1999 “Wages and Unequal Access to Organizational Power: An Empirical Test of Gender Discrimination.” Administrative Science Quarterly, 44: 453-472. Ibarra, H. 1995 “Race, Opportunity, and Diversity of Social Circles in Managerial Networks.” The Academy of Management Journal, 38: 673-703. Johnson N.B., T. A. Scandura. 2004 “The effect of mentorship and sex-role style in male-female earnings.” Industrial Relations, 33:263–274 Kalev, A. 2009 “Cracking the Glass Cages? Restructuring and Ascriptive Inequality at Work.” American Journal of Sociology, 114: 1591–1643. Kalev, A., F. Dobbin, and E. Kelly. 2006 “Best Practices or Best Guesses? Assessing the Efficacy of Corporate Affirmative Action and Diversity Policies.” American Sociological Review, 71:589–617. Kelly, E., and A. Kalev. 2006 “Managing flexible work arrangements in US organizations: formalized discretion or 'a right to ask'.” Socio - Economic Review, 4: 379. Kilbourne, B., P. England, and K. Beron. 1994 “Effects of Individual, Occupational, and Industrial Characteristics on Earnings: Intersections of Race and Gender.” Social Forces, 72: 1149-1176.

Abraham

42

Konrad, A. M. V. Kramer, and S. 2008 “Critical mass: The impact of three or more women on corporate boards.” Organizational Dynamics, 37: 145-164. Kulis, S. 1997 “Gender Segregation among College and University Employees.” Sociology of Education 70: 153-175. Kurtulis, F. A., and D. Tomaskovic-Devey. 2012 “Do Female Top Managers Help Women to Advance? A Panel Study Using EEO-1 Records.” The ANNALS of the American Academy of Political and Social Science, 639: 173–197. Marsden, P. V., and E. H. Gorman. 2001 “Social Networks, Job Changes and Recruitment.” In I. Berg and A. L. Kalleberg (eds.), Sourcebook of Labor Markets: Evolving Structures and Processes: 467-502. New York: Plenum Press. Matsa, D. A., and A. R. Miller. 2011 “Chipping away at the glass ceiling: Gender spillovers in corporate leadership.” American Economic Review, 101: 635–39. McGinn, K.L., and K.L. Milkman. (in press) “Looking up and looking out: Career mobility effects of demographic similarity among professionals.” Organization Science. McPherson, M., L. Smith-Lovin, and J. M. Cook. 2001 "Birds of a Feather: Homophily in Social Networks." Annual Review of Sociology, 27:

415-444. Morrison, A. M., R. P. White, E. Van Velsor, and Center for Creative Leadership. 1987 “Breaking the Glass Ceiling: Can Women Reach the Top of America's Largest Corporations?” Addison Wesley Longman Publishing Co. Nelson, R. L. 1999 “Legalizing Gender Inequality: Courts, Markets, and Unequal Pay for Women in America.” Cambridge, England: Cambridge University Press. Osborne, J. W.

Abraham

43

2000 “Advantages of hierarchical linear modeling.” Practical Assessment, Research, and Evaluation, 7: 1-3.

Penner, A. M., H. J. Toro-Tulla, and M. L. Huffman. 2012 “Do Women Managers Ameliorate Gender Differences in Wages? Evidence from a Large Grocery Retailer.” Sociological Perspectives, 55: 365–381. Petersen, T, and L. A. Morgan. 1995 “Separate and Unequal: Occupation-Establishment Sex Segregation and the Gender Wage Gap.” The American Journal of Sociology 101: 329-365. Petersen, T., and I. Saporta. 2004 “The Opportunity Structure for Discrimination.” The American Journal of Sociology, 109: 852-901. Petersen, T., I. Saporta, and M. Seidel. 2000 “Offering a Job: Meritocracy and Social Networks.” American Journal of Sociology, 106: 763–817. Pfeffer, J. 1991 “Organization Theory and Structural Perspectives on Management.” Journal of Management, 17:789-803. Powell, G. N., and D. A. Butterfield. 1994 “Investigating the "Glass Ceiling" Phenomenon: An Empirical Study of Actual Promotions to Top Management.” The Academy of Management Journal, 37: 68-86. Ragins, B. R., and T. A. Scandura. 1999 “Burden or Blessing? Expected Costs and Benefits of Being a Mentor.” Journal of Organizational Behavior 20:493-509. Reagans, R. 2005 “Preferences, identity, and competition: Predicting tie strength from demographic data.” Management Science, 51: 1374-1383. Reskin, B. F. 2000 “The Proximate Causes of Employment Discrimination." Contemporary Sociology, 29:

319-328. Reskin, B. F., and D. B. McBrier.

Abraham

44

2000 “Why Not Ascription? Organizations' Employment of Male and Female Managers.” American Sociological Review, 65: 210-233. Reskin, B.F. 2003 “Including Mechanisms in Our Models of Ascriptive Inequality.” American Sociological Review, 68: 1–21. Ridgeway, C. L. 1991 “The Social Construction of Status Value: Gender and Other Nominal Characteristics.” Social Forces, 70: 367 – 386. Ridgeway, C. L. 1997 “Interaction and the Conservation of Gender Inequality: Considering Employment.” American Sociological Review, 62: 218–35. Ridgeway, C. L., and S. J. Correll. 2004 “Unpacking the Gender System: A Theoretical Perspective on Gender Beliefs and Social Relations.” Gender and Society 18: 510-531. Ridgeway, C. L., and S. J. Correll. 2006 “Consensus and the Creation of Status Beliefs.” Social Forces, 85: 431–453.

Roth, L. M. 2006 Selling Women Short: Gender and Money on Wall Street. Princeton, N.J.: Princeton University Press. Rothbard, N. P. 2001 “Enriching or Depleting? The Dynamics of Engagement in Work and Family Roles.” Administrative Science Quarterly, 46: 655–684. Rothbard, N. P., K. W. Phillips, and T. L. Dumas. 2005 “Managing Multiple Roles: Work-Family Policies and Individuals’ Desires for Segmentation.” Organization Science, 16: 243–258. Shin, T. 2012 “The Gender Gap in Executive Compensation.” The ANNALS of the American Academy of Political and Social Science, 639: 258–278. Skaggs, S., K. Stainback, and P. Duncan.

Abraham

45

2012 “Shaking Things up or Business as Usual? The Influence of Female Corporate Executives and Board of Directors on Women’s Managerial Representation.” Social Science Research, 41: 936–948. Smith, N., V. Smith, and M. Verner. 2013 “Why Are so Few Females Promoted into CEO and Vice President Positions? Danish Empirical Evidence, 1997-2007”. SSRN Scholarly Paper ID 2256660. Rochester, NY: Social Science Research Network. Stainback, K., and S. Kwon. 2012 “Female Leaders, Organizational Power, and Sex Segregation.” The ANNALS of the American Academy of Political and Social Science, 639: 217–235. Sørensen, J. B. 2000 “The Longitudinal Effect of Group Tenure Composition on Turnover." American Sociological Review, 65: 298-310. Sørensen, J. B. 2007 “Bureaucracy and Entrepreneurship: Workplace Effects on Entrepreneurial Entry." Administrative Science Quarterly 52:387- 412. Sturm, S. 2001 “Second Generation Employment Discrimination: A Structural Approach.” Columbia Law Review 101: 458. Tajfel, H., and J. C. Turner. 1979 “An integrative theory of intergroup conflict.” In W. G. Austin and S. Worchel (eds.), The social psychology of intergroup relations: 33–47). Monterey, CA: Brooks/Cole Tomaskovic-Devey, D., S. Skaggs. 1999 "Workplace Gender and Racial Composition and Productivity: An Establishment Level Test of the Statistical Discrimination Hypothesis." Work and Occupations, 26: 422-445. 1999. Huffman, M. L., and L. Torres. 2002 "It's not only 'Who you Know' That Matters: Gender, Personal Contacts, and Job Lead Quality." Gender & Society 16:793-813. Weber, Max. [1914] 1968 Economy and Society: An Outline of Interpretative Sociology. Translated and edited by Guenther Roth and Claus Wittich, 3 volumes. New York, NY: Bedminster.

Abraham

46

Weeden, K. A. 2005 “Is there a flexiglass ceiling? Flexible work arrangements and wages in the United States.” Social Science Research, 34: 454-482. Wharton, A. S., S. Chivers, and M. Blair-Loy. 2008 “Use of Formal and Informal Work-Family Policies on the Digital Assembly Line.” Work and Occupations, 35: 327-350. Wolf, W. C., and N. D. Fligstein. 1979 “Sexual Stratification: Differences in Power in the Work Setting.” Social Forces, 58: 94–107.

Abraham

47

Figure 1. Branch Positions: Hierarchy and Description of Responsibilities

Non-Managerial Positions “manage self”

Representative

Teller

Officer

Relationship Manager

Executive

Managerial Positions “manage others”

Branch Manager

• Basic products & transactions

• Unsecured products

• Sales & service

• Financial planning & investments

• Advanced sales

• Focus on small business customers

• International products

• Focus on high net worth customers

• Complex products & services

• Leadership • People

management • Business

management & growth

Abraham

48

Composition Hourly WagesPercentage* Mean (SD) Min Max

21.17 12.24 (1.82) 7.50 22.50(101)

14.47 14.89 (1.83) 11.27 20.34(69)

34.59 19.11 (3.37) 13.08 34.01(165)

4.82 22.85 (3.80) 13.44 31.25(23)

9.85 32.17 (7.90) 19.23 53.31(47)

15.09 33.98 (7.55) 21.64 55.82(72)

Overall 100 20.76 (9.01) 7.50 55.82(477)

*Parentheses indicate number of employees

Table 1. Composition of Branch Positions and Mean Hourly Wages (in $) for All Retail Branch Employees, 1999

Account Executive

Relationship Manager

Branch Manager

Teller

Representative

Officer

Abraham

49

Table 2. Basic Descriptive Statistics for Variables of Interest for Managers,1999Overall Female Managers Male Managers

Mean (SD) Percentage Mean (SD) Percentage Mean (SD) Percentage t-testHourly Wages 33.98 (7.55) 33.13 (7.11) 34.69 (7.93)Age (in years) 42.33 (9.02) 44.28 (9.68) 40.68 (8.19)Tenure 10.20 (9.17) 12.72 (8.44) 8.07 (9.32) *Parttime 0.00 0.00 0.00Female 45.83African-American 12.50 12.12 12.82Asian 15.28 15.15 15.38Caucasian 63.89 69.70 58.97Hispanic 8.33 3.03 12.82

N 72 33 39

FN: This table is for 1999, but all patterns are the same for all 4 years.*** p ≤ 0.001, ** p ≤ 0.01, * p ≤ 0.05, ^ p ≤ 0.10 (all two-sided t-tests comparing employees reporting to female versus male managers)

Abraham

50

Overall Female Managers Male ManagersMean (SD) Percentage Mean (SD) Percentage Mean (SD) Percentage t-test

All EmployeesSize 8.12 (2.91) 7.40 (2.24) 8.62 (3.21) ***Female 69.36 71.56 67.82

Female EmployeesHourly Wages 16.64 (4.99) 16.30 (4.39) 16.71 (5.31)Part-time 10.21 13.11 8.02Age (in years) 38.49 (10.33) 38.63 (10.43) 38.05 (10.16)Tenure (in years) 7.68 (6.51) 8.08 (7.44) 6.21 (5.58) *Married 43.66 39.34 46.91African-American 5.99 7.38 4.94Asian 32.75 31.15 33.95Caucasian 42.25 44.26 40.74Hispanic 19.01 17.21 20.37

Male EmployeesHourly Wages 22.55 (9.01) 21.51 (8.60) 22.55 (9.18)Part-time 2.48 4.44 1.32Age (in years) 36.63 (9.95) 38.64 (11.26) 35.15 (8.83) *Tenure (in years) 5.64 (7.41) 6.31 (8.42) 4.96 (6.59)Married 46.28 57.78 39.47 *African-American 8.26 4.44 10.53Asian 22.31 24.44 21.05Caucasian 53.72 55.56 52.63Hispanic 15.70 15.56 15.79

N 405 167 238

FN: This table is for 1999, but all patterns are the same for all 4 years.

Table 3. Basic Descriptive Statistics for Variables of Interest for Non-managerial Employees, by Employee Gender and Manager Gender, 1999

*** p ≤ 0.001, ** p ≤ 0.01, * p ≤ 0.05, ^ p ≤ 0.10 (all two-sided t-tests comparing employees reporting to female versus male managers)

Abraham

51

Table 4. Generalized Estimation Equation Regression Models Predicting Log Hourly Wages of Non-Managerial Employees

Model 1

Model 2

Model 3

Female -0.149***

-0.157***

-0.050*

(0.024)

(0.025)

(0.020)

Manager Female

-0.052***

-0.032^

(0.015)

(0.018)

Manager Female x Female

0.026

0.039^

(0.018)

(0.021)

Representative

0.164***

(0.024)

Officer

0.373***

(0.020)

Executive

0.581***

(0.031)

Relationship Manager

0.791***

(0.032)

Rep x Mgr Female

-0.003

(0.031)

Officer x Mgr Female

0.006

(0.024)

Exec x Mgr Female

0.006

(0.046)

Relation x Mgr Female

-0.043

(0.043)

Rep x Female

-0.009

(0.027)

Officer x Female

-0.030

(0.023)

Exec x Female

-0.058

(0.050)

Relation x Female

-0.077^

(0.044)

Rep x Female x Mgr Female

-0.043

(0.035)

Officer x Female x Mgr Female

-0.024

(0.029)

Exec x Female x Mgr Female

0.036

(0.070)

Relation x Female x Mgr Female

-0.021

Abraham

52

(0.059)

Black -0.066*

-0.069**

-0.017

(0.026)

(0.026)

(0.014)

Asian -0.015

-0.016

0.008

(0.023)

(0.023)

(0.012)

Hispanic -0.103***

-0.103***

-0.018

(0.023)

(0.023)

(0.012)

Age 0.008***

0.008***

0.005***

(0.001)

(0.001)

(0.001)

Tenure 0.016***

0.016***

0.008***

(0.001)

(0.001)

(0.001)

Married 0.211***

0.213***

0.060***

(0.031)

(0.031)

(0.018)

Married x Female -0.196***

-0.200***

-0.051*

(0.037)

(0.037)

(0.021)

Branch Size (num. employees) -0.000

-0.000

-0.001*

(0.000)

(0.000)

(0.000)

Manager Tenure 0.000

0.000

0.000

(0.001)

(0.001)

(0.000)

Constant 2.481***

2.504***

2.325***

(0.038)

(0.038)

(0.025)

N 2,649

2,649

2,649 Number of Groups 1,110

1,110

1,110

Time Periods 4

4

4 Wald Chi2 816.37 ***

845.64 ***

4840.82 ***

DF 17 19 35 Standard errors in parentheses

*** p<0.001, ** p<0.01, * p<0.05, ^ p<0.10

FN: All models include dummy variables for race, controls for employee age, employee age-squared, employee tenure, employee tenure-squared, manager tenure, manager tenure-squared, size of branch, and fixed effects for the state where the branch is located. The omitted category for race is "caucasiian"; for job title is "teller"; for female is "male"; and for manager female is "manager male".

Abraham

53

Table 5. Comparing Predicted Hourly Salary (in $) for Typical Male and Female Tellers by Manager Gender

Female Manager

Male Manager

Female Teller 11.98

11.90

Male Teller 12.12

12.51

Gender Wage Gap

in $ -0.13

-0.61

Fwage as % of Mwage 98.91

95.12

FN: Predicted values for hourly salary from Model 3 (Table 5) for a typical teller profile: a white, single individual, working fulltime, with average tenure and age reporting to a manager with average tenure.

Abraham

54

Abraham

55

Abraham

56

Table 6. Log-Linear Tests of Independence between Joint Distribution of Non-Managerial Employees to Jobs or Parttime and Manager Ge1999 1998 1997 1996

Model Variablesa DFLikelihood Ratio

Chi-Squareb

Likelihood Ratio Chi-Squareb

Likelihood Ratio Chi-Squareb

Likelihood Ratio Chi-Squareb

Model 1: Job Allocation [FB] [M] 9 9.05 12.86 14.82 8.93(0.43) (0.17) (0.10) (0.44)

Model 2: Parttime [FP] [M] 3 4.21 4.51 4.85 1.17(0.24) (0.21) (0.18) (0.76)

N 405 732 948 564

P-values in parentheses.

*** p<0.001, ** p<0.01, * p<0.05, ^ p<0.10

a F = employee female (0 = male, 1 = female), M = manager female (0 = male, 1 = female), B = branch position (i.e. teller, representative, officer, account executive, relationship manager), and P = part-time (0=not parttime, 1 = parttime)b Likelihood Ration Chi-square statistic allows for a test of the null hypothesis that the joint distribution of non-managerial emplyee gender and either job allocation or parttime is not independent of branch manager gender.

Abraham

57

Abraham

58